From discovering a cure for a disease to exploring energy to estimating weather, engineering and science today relies heavily on the simulation and modelling performed on high performance computing systems. But to keep ahead of this exploding demand for computing requires innovation, especially with programming issues and power. New and improved Many-core processors are causing changes in the stable computing architecture used in the past 2 decades. To add on top is the increasing importance of cloud computing which is leading to significant changes in computer system architecture.
Enter Many-core systems
A Multi-core chip is a processing chip with 2 or more identical cores in one package, whereas a Manycore chip is a processing chip that has 10 or more heterogeneous cores in one package. In Manycore system, all cores will share resources including main memory and hard drives. Industry analysts anticipate the advent of affordable “manycore” systems in the next couple of years.
The growth in processing power along with emergence of the Internet, Mobile, and cloud has transformed productivity. Personal computing has fundamentally changed the way companies work, and the way we learn and interact. Instead of drawing processing power by scaling clock speeds, which increases power usage and heat emission to uncontrollable levels, manufacturers are increasing the overall computing power by adding additional CPU cores to the microprocessor package. New processors will come in mixed configurations such as combinations of high- and low-powered cores.
Manycore systems holds the promise of delivering high processing power in the next few years. Developing software applications that harness the full power of manycore systems will be difficult and will require developers to shift from writing serial programs to parallel programs. Apps must scale up or down depending on the capability of the system and should be able to adapt to the changing power and resource availability. Manycore will deliver undeniable experiences in sophisticated image analysis, semantic analysis of data, natural language, computer vision, rendering and manipulation. In the future users will benefit from apps that adapt to the user by observing and learning the user’s behaviour, anticipate his/her computing need, and apply changes to the app’s resource usage.
Final Thoughts:
The shift to manycore paradigm is the next important step in how users will interact with their computer. It renders the potential for the next generation of apps and user experiences. As programmers seek powerful computing platform, the industry continues to deliver in a distributed cluster environment. The next generation of computing is filled with endless possibilities and opportunities.
References:
http://www.hpcwire.com/hpcwire/2007-03-02/our_manycore_future.html
http://software.intel.com/en-us/articles/many-core-processor/
http://en.wikipedia.org/wiki/Multi-core_processor
http://delivery.acm.org/10.1145/1570000/1562783/p56-asanovic.pdf
http://communities.intel.com/servlet/JiveServlet/downloadBody/19480-102-1-22823/Parallel%20programming%20models%20for%20manycore%20systems%20-%20presentation%2C%20Intel%2C%20Wrinn.pdf
http://visualstudiomagazine.com/Articles/2011/07/01/pfven_MultiCore.aspx?cid=sw:social_software_om&Page=1
*If you find something is misleading or not correct then please throw some light on it.